基于声信号的分类器设计及实现  

Design and Realization Classifier Based on Acoustic Signal

作　　者：王建平[1,2] 秦栋泽[1] 焦国太[1] 张可[3] 

机构地区：[1]中北大学机电工程学院,山西太原030051 [2]山西汾西重工有限责任公司,山西太原030027 [3]中国电子科技集团公司第四十七研究所,辽宁沈阳110032

出　　处：《电声技术》2014年第10期53-55,79,共4页Audio Engineering

摘　　要：针对如何利用声信号对目标进行准确识别,以背景信号和车辆信号为研究对象,采用K-NN算法和BP网络算法,对声信号的分类器进行了研究和设计。并以Window XP为仿真平台,Visual C++6.0为开发工具,编写相应程序。仿真结果表明:此算法能有效地对声目标进行分类识别,具有良好的应用性和实用性。In the light of how to use acoustic signal to identify target accurately, based on the background signal and vehicle signal as the research object, using the K nearest neighbour algorithm and Back-propagation neural network algorithm, the classifier of acoustic signal is researched and designed. Corresponding programs are compilied with Window XP as the simulation platform, the Visual C ＋ ＋ 6.0 as development tool. The simulation results show that this algorithm can classificate acoustic target effectively and has good applicability and practicability.

关 键 词：声信号 分类器 K-NN算法 BP算法 

分 类 号：TN911.72[电子电信—通信与信息系统]